Intel® NUCs
Support for Intel® NUC products
Announcements
Do you have improvements you would like us to make on this community site? If so, we would love to hear your feedback! Click here for an 8 question survey. Thanks!

11312 Discussions

Skull Canyon NUC Random Reboots

MA1
Novice
7,401 Views

Hi all,

I've read some past forum posts about this and was wondering if anyone has any other advice. My NUC6i7KYK reboots randomly, sometimes just a few minutes after booting. Other times it is rock stable and never crashes. I have used this machine for months without this problem happening at all, and suddenly it's happening about every day. The strange thing is, it reboots when I'm doing pretty low load tasks like browsing web pages. I was playing Cities skylines at 4K, the machine was blazing hot and fans spinning super fast and it was stable. So it doesn't seem like a heat or load issue.

Specs are: 32gb Patriot 2133 CL14 (2x16gb sodimms), Samsung 950 PRO M.2, steelseries keyboard and mouse plugged into a Startech USB hub which is plugged into the NUC. Using WiFi but no BT at time of reboots. I have an eGPU but started experiencing this issue before I had it. The eGPU was not on at the time of the reboot. Nothing in the NUC is overclocked.

There are no blue screens, nothing in my error logs. When the NUC reboots it goes to an unsuccessful POST message.

This seems to have started since the Creators update of Win 10 but not sure if that's just a coincidence.

To date, my troubleshooting steps have been: update to latest BIOS, reset BIOS to defaults, remove and reseat RAM, try different RAM, reformatted and reinstalled Windows. Nothing has helped.

EDIT: Additionally, I had 2 strange occurrences during reboots. Once after the unsuccessful POST message, it booted to what appeared to be a Windows recovery screen but it was just a blank blue background. The very strange this is that when I typed anything on my keyboard, the screen would flicker. Another time after the unsuccessful POST, it rebooted to a recovery screen asking for my BitLocker key. I put it in but then it said my Windows installation was damaged and it couldn't boot. I just power cycled the NUC and it booted fine.

Is exchanging the NUC the only thing I can do at this point? Any advice is greatly appreciated.

Thanks

Mike

0 Kudos
40 Replies
n_scott_pearson
Super User Retired Employee
899 Views

The most common reason for seeing a spontaneous reset that is then followed by the appearance of an "Unsuccessful POST..." message is a failure in the memory subsystem. That is, one (or both) of the memory buses locked up, the watchdog recovery timer (eventually) resets the system and BIOS POST was then unable to reset/reinitialize the memory buses.

Normally, my suggestion would be to test all combinations with only one SODIMM installed (i.e. SODIMM A in Socket 0, SODIMM A in Socket 1, SODIMM B in Socket 0 and SODIMM B in Socket 1). If this process didn't isolate a bad SODIMM, I would then suggest trying again but with completely different SODIMMs. If the symptom continues to occur, I would conclude that there is an issue at the processor/board level and have the unit replaced.

If you feel you are at the end of the process and want to initiate the RMA replacement process, you need to contact Intel Customer Support directly (you cannot do it here). Here is contact information (by geography):

http://www.intel.com/content/www/us/en/support/contact-support.html# @11 Intel Customer Support Contact Information for US and Canada

http://www.intel.com/content/www/us/en/support/contact-support/emea-contact.html Intel Customer Support Contact Information for Europe, Middle East and Africa

http://www.intel.com/content/www/us/en/support/contact-support/apac-contact.html Intel Customer Support Contact Information for Asia-Pacific

http://www.intel.la/content/www/xl/es/support/contact-support/lar-contact.html Intel Customer Support Contact Information for Latin America

Hope this helps,

...S

MA1
Novice
899 Views

Thank you Scott. Appreciate the reply.

I tried a few combinations of 2 different types of RAM in this NUC and it will always crash/reboot. The one thing though is that neither of these 2 types of memory I have are on the approved/compatible memory list for the NUC. Does that still matter as much now? I know it did in earlier BIOS days. Like I said though, the original memory I had in there (Patriot Viper CL14) was stable for months and months.

The one thing I did not try is a full BIOS recovery, which I just did. Will see how it goes. And to Amy, yes this is BIOS 46 that I'm on. Will let you know if recovery fixes it, otherwise will proceed with RMA exchange of the NUC.

Thanks again

Mike

MA1
Novice
899 Views

BIOS Recovery didn't help. I think I may have narrowed it down to the bottom memory slot. Things seem fine if I have memory just in the top slot. Will continue to test. Intel already contacted me back and are sending a new unit which I should receive tomorrow.

idata
Community Manager
899 Views

/thread/113456 dynamicmange, thank you for the detailed update, and please share your results with the new unit.

Regards,

Amy.

idata
Community Manager
899 Views

Scott, thank you for the detailed input.

dynamicmange, please let me know if you were able to follow the SODIMM combination suggestion.

Just to confirm; is the the BIOS version 046 the one you mentioned as the latest?

Regards,

Amy.

RMcCo2
Beginner
899 Views

I'm having much the same issue.

My specs are:

NUC6i7KYK

BIOS 0046

Samsung 950 PRO 512 SSD

16GB Ram - G.Skill DDR4-2133 8GB single channel Ripjaws SODIMM ( x 2)

Windows 10 64-bit, Ubuntu 16.04

The random reboots were initially rare, but in the past 6-8 weeks they are occurring daily, sometimes multiple times a day. I found that updating the firmware for the 950 PRO has reduced the frequency of the reboots, but they still occur. I don't get any messages on reboot.

n_scott_pearson
Super User Retired Employee
899 Views

Ryan,

Are these resets occurring in both Windows and Linux or only while in Windows?

...S

RMcCo2
Beginner
899 Views

Thanks Scott - both Windows and Linux experience the random reboots.

The common thread with these issues seems to be the SSD.

MA1
Novice
899 Views

Yeah I'm kind of at a loss here. I got my replacement NUC, loaded a fresh clean copy of Windows 10. Got all of my drivers loaded, everything was going great, and then bam.. reboot. So obviously my old NUC was not bad. Really the only remaining possibility for me is the NVM-e drive, or maybe some recent BIOS update causing incompatibility with it. I know for a fact this started the day after I installed the Win 10 Creators update and I actually just came online to download the old version and try to reinstall it. But I see that you're having crashes on Linux so I doubt this is an OS issue.

Definitely have a common thread of the 950 pro. I am on the latest firmware, which did not help or reduce crashes. Magician says the drive is good and the benchmark numbers are great. At this point I am just experimenting, but I am going to roll back to the BIOS before 42 and see if that helps. This system used to be so stable that I would leave it on all the time and never had crashes. The NUC was rock solid for months. I used it as a remote desktop server. It had 2 Hyper-V VM's running on it all the time keeping tunnels open and doing various things. I used to always be able to reliably walk up to the computer, log in and use it even if I hadnt rebooted in weeks. Now it crashes all the time, literally every time I boot it up, it will crash whether its in 10 minutes or 2 hours. It is not stable at all anymore.

I guess I will try to contact Samsung on Monday and see what they think.

MA1
Novice
899 Views

Contacted Samsung. I have to send the drive in to them for testing and if it's bad they'll repair or replace. I think what I'll try to do to eliminate that as a possibility is install an OS on my external 256gb usb3 flash drive and use that for a while and see if I get reboots. Fun times.

idata
Community Manager
899 Views

/thread/113456 dynamicmange, my apologies for the inconvenience. You basically narrowed this down to the SSD, and if there is chance to test a different one to confirm this is a good idea.

Please let me know how it goes with the your other drive.

Regards,

Amy.

MA1
Novice
899 Views

Unfortunately it's not very easy for me to test another M.2 drive. The only other one I have is in my working laptop which I need to use since my NUC is so unstable.

To be honest I am not 100% convinced it is the NVM drive. I feel like I would see something more along the lines of a Windows BSOD or something like that rather than a random reboot. Of course I am just guessing here.

I am still experimenting a little before I send my drive in to Samsung. I have rolled back to BIOS 042 because I seem to remember installing 045 and 046 very close together right when I started having issues, so I want to eliminate that possibility. So far the system has been stable all day after going back to 042. And as I said I will get an OS (most likely Ubuntu) running on a Flash drive for a while so I'm not even using the Samsung.

Will report back.

Thanks

n_scott_pearson
Super User Retired Employee
899 Views

No, not necessarily. If the issue is manifesting as a PCIe bus lockup (which will hang the whole system), the system will be hard reset by the watchdog timer. There would be no opportunity for Windows to detect or do anything about it (like log something)...

...S

MA1
Novice
899 Views

Gotcha. Thanks. Fingers crossed but so far I haven't had a reboot yet today after rolling back to BIOS 042. I'm probably jynxing myself here but it has definitely been the longest block of time without a reboot in the last few days. I won't really think this is a fix until I have a couple days up time with no crash. I will report back.

Thanks again

RMcCo2
Beginner
899 Views

I also tried rolling back to BIOS 0042, but unfortunately the reboot issue persists.

I'll try to test a SATA M.2 drive over the weekend and report back.

MA1
Novice
899 Views

Interesting. I went an entire day without a reboot yesterday. First time that has happened since I started having this problem. Will continue to monitor and test. When you rolled back to 042 what method did you use? I did a power button recovery then reset to defaults afterwards.

idata
Community Manager
899 Views

/thread/113456 dynamicmange, keep us posted.

Regards,

Amy.

MA1
Novice
899 Views

It has been several days since I have had the random reboots. At this point I have to say that rolling back to BIOS 042 via recovery method, then resetting to defaults and setting everything up from scratch fixed my issue. Is it possible that BIOS 045 or 046 introduced some sort of incompatibility with my hardware?

RMcCo2
Beginner
899 Views

Good to hear things have been stable. Definitely post back if the reboots occur again.

I have just done a fresh Windows 10 install on a Samsung SM951 (AHCI version) and put the BIOS back to 0046. I will see how this runs over the next week.

idata
Community Manager
524 Views

/thread/113456 dynamicmange, thanks for the update.

At this stage we have not received reports of these BIOS versions causing any type of issues. It is possible but this could be an isolated case, since version BIOS 0042 is working I recommend sticking with that version and wait for future BIOS releases.

Regards,

Amy.

Reply